In spite of significant improvements in antibiotic therapy and in intensive care, sepsis, and its sequelae, sepsis syndrome or septic shock (collectively, Asepsis@), remain a leading cause of morbidity and mortality among hospitalized patients. Sepsis is triggered by gram-negative and gram-positive bacteria, fungi, and other pathogenic microorganisms. These organisms release toxins at the nidus of injury or infection, that in turn trigger the release of cytokines and other mediators. If infection is not controlled, endotoxin and/or other mediators of inflammation may enter the circulation, initiating sepsis and the cascade of events that leads to endothelial damage, hypotension and multi-organ failure. Gram-negative bacteria are responsible for a large number of such episodes, which are associated with a high mortality rate. See, e.g., Centers for Disease Control, “Increase in national hospital discharge survey rates for septicemia B United States, 1979-1987,” Morbid, Mortal, Weekly Reports, 39, 31 (1990). In patients who develop septic shock caused by gram-negative bacteria, the fatality rate may reach 50% or more. See, R. C. Bone et al., N. Eng. J. Med., 317, 653 (1987). Eseherichia coil remains the leading causative organism, accounting for 40 to 52% of gram-negative blood isolates (S. Chamberland et al., Clin. Infect. Dis., 15, 615 (1992); B. E. Kreger et al., Am. J. Med., 68, 332 (1980)).
Lipopolysaccharide (LPS, endotoxin) is the major component of the outer membrane of gram-negative bacteria and is responsible for many of the pathophysiological effects observed during infections with gram-negative pathogens that may lead to septic shock and death (E. T. Rietschel et al., Scient. Amer., 267, 54 (1992); FASEB J., 8, 217 (1994)). Enterobacterial LPS consists of three domains, i.e., lipid A, core region and O-specific chain, of which lipid A is structurally the most conserved among different pathogenic bacteria, and represents the toxic principle of LPS (C. A. H. Raetz, Ann. Rev. Biochem., 59, 129 (1990); E. T. Reitschel et al., Infect. Dis. Clin. North Am., 5, 753 (1991); C. Galanos et al., Eur. J. Biochem., 148, 1 (1985)). The structure of E. coil J5 LPS is shown in FIG. 1 (from Galanos et al. (1985)). As the toxic effects exerted by LPS are independent of the viability of bacteria and considering the increasing resistance of pathogenic bacteria to antibiotics, the search for alternative treatment strategies for sepsis is of major importance.
One of the most promising approaches for the immunotherapy of sepsis is passive immunization with antibodies that are directed against the conserved regions of LPS, i.e., lipid A and the core region. Such antibodies are expected to be cross-reactive with different gram-negative pathogens and might therefore be cross-protective. Passive immunization with polyclonal or monoclonal antibodies (Mabs) against bacterial LPS has shown protective effects in animal models of sepsis. It was shown that partially detoxified LPS from E. coli J5 could elicit polyclonal antibodies in rabbits that provided passive protection against Pseudomonas aeruginosa infections in rats (A. K. Bhattacharjee et al., J. Infect. Dis., 170, 622 (1994)). Similarly, it has been shown that monoclonal antibodies against E. coil J5 could provide passive immune protection against heterologous bacteria challenges in mice (M. P. Schutze et al., J. Immunol., 142, 2635 (1989)). See also, F. E. DiPadova et al., Infect. Immun., 61, 3869 (1993); J. D. Baumgartner et al., Immunobiology, 187, 464 (1993). However, protection generally requires that the antibodies (Ab) be administered before sepsis pathology begins. This indicates that passive immunization has the potential to provide prophylactic protection but not therapeutic efficacy.
Prophylactic protection is best provided by active immunization, or vaccination, rather than by passive immunization. The induction of protective antibodies could potentially be achieved by immunization with LPS presented in an appropriately modified form or via mutant bacteria defective in LPS biosynthesis (rough mutants) (C. Galanos et al., Eur. J. Biochem., 31, 230 (1972); S. C. Bruins et al., Infect. Immun., 17, 16 (1977)). Escherichia coli J-5, a rough mutant of E. coli O111:B4, has been used in the majority of immunological studies for more than three decades in an attempt to induce broadly cross-reactive and cross-protective antibodies directed against LPS. In fact, immunization of mice with heat-killed E. coli J5 cells can elicit active immune protection against a challenge of the mice with Haemophilus influenzae type b (M. I. Marks et al., J. Clin. Invest, 69, 742 (1982)). See also, J. B. Baumgartner et al., J. Infect. Dis., 163, 769 (1991). Multiple injections of purified, detoxified E. coli J5 LPS can also function as an antigen to elicit cross-protective anti-LPS Abs. A. K. Bhattacharjee et al., J. Infect. Dis., 173, 1157 (1996) prepared a noncovalent vaccine using partially detoxified J5 LPS and the outer membrane protein of N. meningitidis Group B.
However, development of a safe and efficacious vaccine against sepsis is hindered by problems associated with the preparation of non-toxic LPS antigens that can elicit cross-protective antibodies to many kinds of bacteria. As shown in FIG. 1, the diglucosamine moiety of LPS is substituted with ester-linked phosphates, ester- and amide-linked fatty acids and with glycosidically linked polysaccharide (C. R. Raetz, Annu. Rev. Biochem., 59, 129 (1990)). The non-lipid parts of the LPS molecule contain epitopes that can participate in eliciting beneficial antibodies; and the lipid (or fatty acid) substituents contain determinants of LPS toxicity (C. Galanos et al., Eur. J. Biochem., 148, 1 (1985); T. Reitschel et al., Infect. Dis. Clin. North Amer., 5, 753 (1991)). Thus, to detoxify LPS, attempts have been made to hydrolytically remove fatty acids while minimizing the loss of other epitopes. One approach uses mild alkaline hydrolysis that releases ester-linked fatty acids from the diglucosamine backbone. The problem with this method is that it does not release amide-linked fatty acids, and so does not provide for complete detoxification. In the case where this treatment was applied to LPS from E. coli J5, the partial deacylation of LPS diminished LPS pyrogenicity about 100 fold (A. K. Bhattacharjee et al., J. Infect. Dis., 170, 622 (1994)). However, the partially deacylated product still exhibited pyrogenic activity at a dose lower than the dose needed to elicit protective antibodies.
The other approach for detoxification of LPS uses mild acid hydrolysis. This approach provides for greater attenuation of toxicity but causes more extensive destruction of polysaccharide epitopes. This treatment cleaves the glycosidic bond between the inner core of LPS and the lipid A diglucosamine backbone (S. J. Cryz et al. (U.S. Pat. No. 5,370,872); R. K. Gupta et al., Infect. Immunol., 63, 2805 (1995); C. Galanos et al., Eur. J. Biochem., 148, 1 (1985)). After hydrolysis, the polysaccharide fraction is collected for use as antigen, and the diglucosamine with attached fatty acids and phosphates is discarded. The problem with this method is that acid hydrolysis removes epitopes associated with the diglucosamine, and also partially modifies the structure of LPS polysaccharides. In the case of E. coli J5 LPS, mild acid hydrolysis treatment can generate polysaccharide antigens that are missing both sugar groups and phosphate groups known to be present in the polysaccharide core of native LPS. Thus, in addition to the absence of the diglucosamine backbone, the detoxified LPS polysaccharides would be depleted of ethanolamine phosphate and non-reducing terminal 3-deoxy-manno-oct-2-ulosonic acid (KDO) residues (S. Muller-Loennies et al., Eur. J. Biochem., 260, 235 (1999)).
The preparation of vaccines based on detoxified LPS is also hampered by problems associated with the preparation of a suitable carrier protein for LPS antigens. A carrier protein is required because LPS polysaccharides do not have epitopes that activate helper T-cells, and without a carrier, they do not induce immune memory that is needed to elicit high titers of long-lived antibodies (J. B. Robbins et al., J. Infect. Dis., 161, 821 (1990)). Detoxified bacterial toxins, such as tetanus toxin or Toxin A, referred to as “toxoids” have been used as carriers for polysaccharide antigens. When covalently linked to a carrier protein, detoxified LPS polysaccharides function as haptens and some immunogenic properties of the carrier are conferred to the linked polysaccharides. In particular, T-cell epitopes in the carrier can induce immune memory responses to the linked polysaccharide haptens.
A limitation in the use of toxoid carriers is that toxoids can cause carrier-specific epitopic suppression of haptens. In experimental animals, this phenomenon occurs when animals are immunized against a toxoid before they are vaccinated with toxoid-hapten conjugate (C. Berquist et al., Infect. Immun., 65, 1579 (1997); L. A. Herzenberg et al., Nature, 285, 664 (1980); M. P. Schutze et al., J. Immunol., 135, 2319 (1985)). There is evidence that acquired immunity to a toxoid can also cause carrier-specific epitopic suppression in humans (D. DiJohn et al., Lancet, 2, 1415 (1989)). Adult humans would be more likely to have immunity to toxoids than young children due to increased probability of exposure. This observation leads to a prediction that toxoid-polysaccharide conjugate vaccines would be less efficacious in adults than in young children.
Therefore, a continuing need exists for immunogenic conjugates that can provide protection against gram-negative sepsis in mammals susceptible thereto.